Ensemble empirical mode decomposition (EEMD) and Teager-Kaiser energy operator (TKEO) based damage identification of roller bearings using one-class support vector machine.

Ensemble empirical mode decomposition (EEMD) is a newly developed noise assisted method aimed to solve mode mixing problem exists in empirical mode decomposition (EMD) method. Although EEMD has been utilized in various applications successfully, small defects of bearings are not able to be detected, especially in automatic defect detection, when only healthy samples are available for training. Teager-Kaiser energy operator (TKEO) technique is a non-linear operator that can track the energy and identify the instantaneous frequencies and instantaneous amplitudes of signals at any instant. As Teager-Kaiser energy operator (TKEO) technique detects a sudden change of the energy stream without any priori assumption of the data structure, it can be utilized for vibration based condition monitoring (non-stationary signals). In this study it is investigated whether an automatic method is able to diagnose a small defect level of roller bearings through processing of the acquired signals. After applying TKEO on IMFs decomposed by means of EEMD, the extracted informative feature vectors of the healthy bearing are used to construct the separating hyperplane using one-class support vector machine (SVM). Then, success rates of state identification of both samples (healthy and faulty) are examined by labelling the samples. The data were generated by means of a test rig assembled in the labs of the Dynamics & Identification Research Group (DIRG) at mechanical and aerospace engineering department, Politecnico di Torino. Various operating conditions (three shaft speeds, three external loads and one small size damage on a roller) were considered to obtain reliable results.